Many psychometricians and behavioral geneticists believe that high heritability of IQ test scores within racial groups coupled with environmental hypotheses failing to account for the differences between the mean scores for groups lends plausibility to explanations of mean differences in terms of genetic factors. I show that heritability estimates and the statistical analysis of variance (AOV) on which they are based have limited relevance in exposing genetic and environmental factors operating within any single group or population. I begin with agricultural investigations, where replication of genetic types and control over environmental factors are possible, and highlight the difficulties of moving from AOV of observed traits to investigation of measurable genetic factors. The difficulties can only be exacerbated for human data sets, which are equivalent to a crop trial in which each variety is replicated in only one or two of the locations.
The possible heterogeneity of factors is an important consideration here. In analyzing crop trials, the AOV is most useful for generating hypotheses about measurable factors when genetic varieties can be grouped so as to make it more plausible that the factors underlying the similar responses of varieties within a group are homogeneous. Even then, this does not imply that in other groups the same set of genetic factors are being modulated by the same set of environmental factors to produce the varieties’ responses, or, by extension, the same set of factors produces the differences between the mean responses for groups. This article and its companion (Taylor in press 2006 Biological Theory 1[4]) seek to open more conceptual space for the challenge of deriving empirically validated models of developmental pathways whose components are heterogeneous and differ among individuals at any one time and over generations.